NASA MODIS Characterization Support Team (MCST) Geometric ...€¦ · Scan-to-scan underlap for...

(Terra, Aqua) MODIS Geolocation Status

NASA MODIS Characterization Support Team (MCST) Geometric Calibration Group

Guoqing (Gary) Lin, SSAI/GSFC Code 619

Robert E. Wolfe, NASA/GSFC Code 619

James C. Tilton, NASA/GSFC Code 606

Ping Zhang, Bin Tan, SSAI/GSFC Code 619

John Dellomo, GST/GSFC Code 619

NASA MODIS-VIIRS Science Team MeetingCalibration Workshop

Silver Spring, MD10/18/2018

Outline

Lin et al., 18 Oct 2018 2

• Geolocation performance for MODIS on Terra and Aqua

– Overall performance

– Trends & details

• Added Info -- scan-to-scan underlaps & on-orbit focal length

• Future work

• Conclusions

Overall Geolocation performance

Residuals Terra C6 Aqua C6 Terra C6.1 Aqua C6.1 NPP VIIRS C1

Track mean 1m 3 m 0 m 1 m 12 m

Scan mean 0 m 1 m 0 m 0 m 4 m

Track RMSE 43 m 47 m 43 m 46 m 58 m

Scan RMSE 44 m 54 m 44 m 53 m 52 m

Data-days 6739 (18.4 yrs) 5917 (16.2 yrs) 6725 (18.4 yrs) 5910 (16.2 yrs) 2447 (6.7 yrs)

Missing days 61 10 59 10 1

Daily matched

GCPs w/ B1/I1

258 190 258 223 204

• Nadir equivalent accuracy (RMSE = Root Mean Square Error)

– Mostly within 20% band B1 HSI (250 m) = 50 m @ nadir (75 m for VIIRS I1);

– Within 10 % for HKM bands and 5% for KM bands

• Band-to-band mis-registration to other bands adds bias to RMSE :

• Other features for MODIS geolocation

– Aqua uses definitive ephemeris data 27 hour latency (Terra uses TDRSS-based on-board ephemeris)

– Aqua C6.1 corrected pointing variations (most of them) caused by AMSR_E stop -go slow – full stop activities

Lin et al., 18 Oct 2018

22 RMSE

3

Terra trend and update details

-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)

Years since Jan. 1, 2000

Daily 16-day Global 16-day Southern Hemisphere 16-day Northern Hemisphere

-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)



Terra C6.1 long-term trend (uncorrected)

RMSE with no correction: Track: 49 m (+6 m vs C6) Scan: 48 m (+4 m vs C6)

-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)



-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)



Actual Terra C6.1 residuals

C6.1 RMSE Track: 43 m Scan: 44 m, nadir equivalent

±20% 250m-band pixel size

-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)



-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)



Actual Terra C6 residuals

C6 RMSE Track: 43 m Scan: 44 m, nadir equivalent


Aqua trend and update details

-50

-25

0

25

50

75

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)



-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)



Aqua C6.1 Long-term Trend (unorrected)

RMSE with no correction: Track: 71 m (+25 m vs C6.1) Scan: 55 m (+2 m vs C6.1)

AMSR-E, 2->0 rpm

2015-12-4T5:34zAMSR-E, 40->0 rpm

2011-10-4T7:27z

AMSR-E, 0->2 rpm

2013-12-5, 16:50 UTC

-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)



-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)

.


Lin et al., 18 Oct 2018 10

Actual Aqua C6.1 residuals

C6.1 RMSE Track: 46 m, Scan: 53 m, nadir equivalent


-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Tra

ck (

ad

j.)

res.

(m)



-75

-50

-25

0

25

50

75

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Scan

(ad

j.)

res.

(m)


Lin et al., 18 Oct 2018 11

Actual Aqua C6 residuals

C6 RMSE Track: 46 m, Scan: 54 m, nadir equivalent

AMSR-E, 2->0 rpm

2015-12-4AMSR-E, 40->0 rpm

2011-10-4±20% 250m-band pixel sizeAMSR-E, 0->2 rpm

2013-12-5

Leap-second handling

Lin et al., 18 Oct 2018 12https://en.wikipedia.org/wiki/Leap_second

year June30

Dec 31

Terra C6

Terra C6.1

Aqua C6

Aqua C6.1

2005 +1 good good good good


2012 +1 good good Note 1 Note 2



Notes 1 & 2:

A error was recently found and corrected in the Aqua spacecraft ephemeris and attitude

dataset for the first 12 hours from 00z to 12 z on 7/1/2012 after leap-second insertion at the

start of the day (2012-06-30T23:59:60z). Data products have been replaced in Sept. 2018

Caution: geolocation errors may still be large after leap-second in the first 5-min granule:

2006-01-01T00:00-00:05

2009-01-01T00:00-00:05

2012-07-01T00:00-00:05

2015-07-01T00:00-00:05

2017-01-01T00:00-00:05

We may “hide” the granules

upon further examination

Scan-to-scan underlap

Lin et al., 18 Oct 2018 13

where

n = # detectors,

p = detector “pitch” interval in the track direction,

F = effective focal length (EFL),

T = scan period, tied to F by BBR requirements,

i = inclination angle, > 90o for MODIS and VIIRS,

VECI = spacecraft ground speed in the inertial frame,

h = range from instrument to earth terrain surface

Vearth0 = surface speed of earth rotation at equator,

Overlap < 0 underlap.

𝑂𝑣𝑒𝑟𝑙𝑎𝑝 = 𝑛𝑝

𝐹𝒉 − [𝑽𝑬𝑪𝑰 − 𝑉𝑒𝑎𝑟𝑡ℎ0 𝐜𝐨𝐬 𝒊 ]𝑇

This term is added for

JPSS-3&4 VIIRS

Instrument properties

Earth property

Satellite properties

Scan-to-scan underlap for Terra MODIS

Lin et al., 18 Oct 2018 14

-90

-75

-60

-45

-30

-15

0

15

30

45

60

75

90

600

700

800

900

1000

1100

1200

39830 44830 49830 54830 59830 64830 69830 74830 79830

Lati

tud

e (

de

g)

Alo

ng

Tra

ck G

SD

(m

)

Scan line #

MODIS Terra 2016-06-30 (one orbit)

Within Scan

From Odd scan to Even Scan

From Even scan to Odd scan

Naidr Latitude

underlap

overlap

• Terra MODIS has underlap of ~ 90 m around 15oN in all neighboring scans

• High terrain worsens the underlap where it occurs

• The altitude variations, minimal near equator and higher near poles, can be seen from the variations of within scan ground sampling distance (GSD)

Terrain effects

Terrain effects

Scan-to-scan underlap for Aqua MODIS

• Aqua MODIS has more serious underlap in alternating scans around 15oN

• Orbit asymmetry due to drifts makes underlap more over descending side (~ 170m) than the ascending side (~ 150 m) in this case, may reverse in another caseLin et al., 18 Oct 2018 15

-90

-75

-60

-45

-30

-15

0

15

30

45

60

75

90

600

700

800

900

1000

1100

1200

531 5531 10531 15531 20531 25531 30531 35531 40531

Lati

tud

e (d

eg)

Alo

ng

Tra

ck G

SD (

m)

Scan line #

MODIS aqua 2016-06-30 (one orbit)

Within Scan

From Odd scan to Even Scan

From Even scan to Odd scan

Latitude

overlap

underlap underlap

Terrain effects

Terrain effects

Aqua MODIS underlap vs scan sample

• Maximum underlap, if exists, exists at nadir and around 15oN

• Due to off-nadir bow-tie effects, scan-to-scan underlap closes off-nadir ~ 122 km for the large one and ~ 60 km for the smaller one in this case.

Lin et al., 18 Oct 2018 16

900

950

1000

1050

1100

1150

1200

-150 -100 -50 0 50 100 150

Gro

un

d S

am

ple

Dis

tan

ce (m

)

Sample # off-nadir

Aqua MODIS underlap off-nadir

Within scan

From odd scan to even scan

From even scan to odd scan

overlap

underlapoverlap

Underlap, max 170 m

-1

-0.5

0

0.5

1

B4

B4

B4

B4

B4

B4

B12

B12

B12

B12

B12

B2

B2

B2

B2

B2

B2

B7

B7

B7

B7

B7

B7

B32

B32

B32

B32

B32

B32

EFL

dev

iati

on

(% N

om

ina

l)

Band Name

EFL on-orbit measurements for Terra MODIS

-1

-0.5

0

0.5

1

B4

B4

B4

B4

B4

B4

B4

B4

B12

B12

B12

B12

B12

B12

B12

B12

B2

B2

B2

B2

B2

B2

B2

B2

B7

B7

B7

B7

B7

B7

B7

B7

B32

B32

B32

B32

B32

B32

B32

B32

EFL

dev

iati

on

(% N

om

ina

l)

Band Name

EFL on-orbit measurements for Aqua MODIS

On-orbit focal length measurements

• On-orbit measured effective focal length (EFL) varies among builds and bands.

• A +0.1% EFL change means +10 m change in scan-to-scan underlap where it exits.– Band B7 has less underlaps for both Aqua and Terra MODIS

– Band B12 has more underlaps for Terra MODIS

Lin et al., 18 Oct 2018 17

Vis FPA NIR S/MWIR LWIR

LWIR

NIR S/MWIR

Vis FPA

2002 2018

2002 2018

Future work (C7.0 soon?)

Lin et al., 18 Oct 2018 18

1) Routine monitor and LUTs update as needed

2) Refresh ground control point chip library

1) Chips are extended from 24x24 to 36x36 km

2) Error measurements are extended from ± 45o to ± 55o)

3) Automate GEO LUT updates

4) Update DEM, LWM (year by year?)

5) Add 250m offsets(?) to geolocation files (currently 1000m with 500m offsets)

6) Create Level-1 geolocation web (needed?)

1) similar to L1B

Anything Else?

Any change in priority order above

Concluding Remarks

Lin et al., 18 Oct 2018 19

• Geolocation performance for MODIS on Terra and Aqua is

good

– mean errors for band B1 near 0 and uncertainties are ~ 50 m at nadir for

the missions, statistically

– C6.1 corrected artifacts in C6 caused by LUT updates

– AMSR_E stop-go-stop activities on Aqua induced MODIS geolocation

errors, which were corrected in C6.1

– Scan-to-scan underlaps exist around nadir in the equator regions

• Maximal underlaps are at nadir near 15oN, at ~ 150 m for Aqua MODIS B2

Backup slides: sun angle dependent Residuals

Lin et al., 18 Oct 2018 20

y = -0.0003x2 - 0.0076x + 4.4306R² = 0.2968

y = -0.0025x2 + 0.5511x - 22.682R² = 0.7863, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

-20 0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Track

y = 0.0002x2 - 0.0574x + 3.4522R² = 0.0981

y = -0.0031x2 + 0.7185x - 34.362R² = 0.8647, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

-20 0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Scan

y = -0.0003x2 - 0.0116x + 3.12R² = 0.3189

y = -0.0025x2 + 0.5511x - 22.682R² = 0.7863, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

-20 0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Track

y = 0.0001x2 - 0.0561x + 3.7938R² = 0.1107

y = -0.0031x2 + 0.7185x - 34.362R² = 0.8647, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

-20 0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Scan

Lin et al., 18 Oct 2018 21

Terra Sun angle CorrectionC6.1 resultsC6 results

y = -0.0021x2 + 0.3683x - 11.3785R² = 0.5607

y = -0.0043x2 + 0.6999x - 19.695R² = 0.7678, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Scan

y = 0.0009x2 - 0.0951x + 2.22R² = 0.4793

y = 0.0026x2 - 0.3546x + 4.3033R² = 0.7477, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Track y = 0.0008x2 - 0.0841x + 0.0915R² = 0.4656

y = 0.0026x2 - 0.3546x + 4.3033R² = 0.7477, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Track

y = -0.0020x2 + 0.3559x - 12.2284R² = 0.5040

y = -0.0043x2 + 0.6999x - 19.695R² = 0.7678, C5 used for C6, C6.1

-100

-75

-50

-25

0

25

50

75

100

0 20 40 60 80 100 120 140 160 180

Sh

ift (m

ete

rs)

Sun angle (degrees)

Scan

Lin et al., 18 Oct 2018 22

Aqua Sun angle CorrectionC6 results C6.1 results

Root cause – example by VIIRS

• Earth rotation casts a component of speed in the VIIRS track & satellite inclination direction

• A sensor Field of regard (FOR) in the track direction FOR = n ASI h needs to cover all the

time everywhere the scan-to-scan travel distance, i.e. FOR >= Ds2s=T[VECI -Vearth0cos(i)] . Otherwise, scan-to-scan underlap occurs

• For details, see Lin et al (2016),“Trending of SNPP ephemeris and its implications on VIIRS geometric performance”, Proc. of SPIE, Vol. 9972, 99721K, doi: 10.1117/12.2239043.

Lin et al., 18 Oct 2018 23

i= 98.67o

VECI = 6,584 m/s

- Vearth (in ECI) = 464 m/s

VECR = 6,670 m/s

SSP1(to)

SSP2

(to +1s)

Equator

VRTA = 3,000 to 18,000 km/s, perpendicular to VECI.16/32 detectors registeron the ground in parallel with VECI

simultaneously

E

- Ve_track = 70 m/s

- Vearth = 464 m/s

NASA MODIS Characterization Support Team (MCST) Geometric ...€¦ · Scan-to-scan underlap for...

Documents

Transcript of NASA MODIS Characterization Support Team (MCST) Geometric ...€¦ · Scan-to-scan underlap for...